Shed diff report #305
Shed diff report #305
Conversation
By using the to_json filter in jinja, we don't need to manually render to a json string
|
Ping @martenson, this will make the shed tests much easier to read. :) |
|
Will get failing test cases figured out presently. |
|
Let me know if you need help with the test cases, otherwise this looks very good to me. Thanks @erasche. |
|
@jmchilton got my issue figured out :) Also adding test cases for new shed_diff xunit report. Will push shortly. |
|
Okay, I'm at a loss for why my tests are still failing when they aren't locally. I'll attack this some more tomorrow. |
|
Yeah, they work locally for me too. Pushed a small tweak to the test cases that should improve the assertion error message reporting. |
|
(TIL you can run individual tox tests with |
@jmchilton apparently os.walk is nondeterministic. I believe This was causing the repositories to be discovered in varying orders on Travis's machines (as opposed to ours), and thus when the XML results were compared to the static files, cat2 would be before cat1, *sometimes*. From https://docs.python.org/2/library/os.html#os.walk > impose a specific order of visiting found via http://stackoverflow.com/a/9062668
hexylena
force-pushed
the
shed-diff-report
branch
from
220b2c1
to
452654c
Compare
|
It finally passes! Whuuuu! @jmchilton haha. Sorry, just been wrestling with this for days and to find out it was all because of stupid non-deterministic library methods... |
|
Brilliant, thanks a bunch @erasche. |
This produces a (fake) XUnit report for the shed_diff command. It's not complete as-is, but it'll provide essential functionality immediately. The XUnit reports will tell you if a shed diff returned a repository as being identical, different (failure), or if the diff command failed (error). It will not show a copy of the diff as I couldn't figure out how to cleanly extract that.
@jmchilton if you have suggestions for how this should be better implemented, please let me know. I see that most of the commands are pretty small and most of that functionality is tucked away in
planemo.shedand friends, but here I went with the implementation I was comfortable with. If you have suggestions how I should get at the diff output I'd be happy to make another PR with those (or update this one).(I branched off of that branch, I didn't want to deal with merge issues since these both needed common changes to the code, blargh.)